Facing the Unknown
One use of a robust control system is to ensure that a satellite will reach its orbit even in the event unknown factors force it off its planned trajectory. CSL's control work in the 1990s zeroed in on robust control theory, which deals with how to design controllers that maintain a system's stability, despite unknowns in the model.
CSL successfully applied a game-theoretic approach to robust control.
In this approach, the control problem is looked at as a two-player game: the controller versus uncertainty. The goal is for the controller to safeguard the achievable performance from the worst that can happen.
CSL developed the new concept of cost-to-come, a counterpart to cost-to-go. This theory was widely used to control nonlinear systems in which measurements are corrupted.
Fast and Slow Dynamics
CSL developed a comprehensive decomposition and aggregation theory for analyzing and synthesizing controllers in large-scale, uncertain systems that exhibit fast and slow dynamics.
Optimization with Constraints
New necessary and sufficient conditions and new efficient global optimization methods solved problems in NASA Mars planning, financial engineering, and image processing.
CSL co-developed a smart ice-management system for airplanes, which senses in-flight ice buildup on an airplane. The system alerts the pilot, prevents the plane from making dangerous maneuvers, and adapts the flight control system to icing conditions.
Ice buildup can pose serious danger because it affects an airplane's aerodynamics. Also, when pilots lack crucial information about icing, they sometimes make misinformed decisions.